INTRAMOLECULAR PHOTOCHEMICAL CYCLOADDITION REACTIONS O F NONCONJUGATED OLEFINS WENDELL L. DILLING
Edgar C. Britton Research Laboratory, The Dow Chemical Company, Midland, Michigan 48640 Received January 31, 1966 CONTENTS
I. Introduction and Scope of the Review.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Cycloadditions Involving Carbon 1,2 Addition. .................................. A. Acyclic Compounds.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Olefins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. a,@-UnsaturatedCarbonyl Compounds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Monocyclic Compounds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
373 374
..................................... .................................
380
376 379
. . . . . . . . . . . . . . . . . . . 384
E. Tetracyclic C IV. Cycloadditions Involving Carbon 1,4 Addition. . . . . . . . . . . . . . . . . . .
VI. References.. . . . . . . . . . . . . . . . . . .
I. INTRODUCTION AND SCOPE OF
THE
............
REVIEW
A
a [c]
(Eq5)
__v_
-%
The photochemica~cycloaddition of adjacent pyrimidine rings in polynucleotide chains, such as DNA, to give substituted cyclobutanes is an intramolecular addition which would fall under the scope of this review. However, these reactions have been reviewed recently (65, 103) and therefore will not be discussed in this review. The review does not include cyclobutene formation from conjugated olefins, a process which could be considered an intramolecular cycloaddition (Eq 6).
(Eq
(Eq2)
I
392
photochemical cycloaddition reactions involving two such additions have been reported (Eq 5 ) . The second reaction probably involves an intramolecular addition of an intermediate formed from the initial cycloaddition (57). These reactions are
This review is intended to cover all work, except as noted below, published through 1964 and most of 1965 on photochemical cycloaddition reactions which intramolecularly with nonconjugated olefins, Le., molecules in which the two addend groups are not conjugated in the classical sense (Eq 1-3). These reactions may be 1,2 or 1,4 additions or a combination of the two. So far as is known no 1,6 or higher intramolecular photochemical addition types have been reported'
-
...........................
(Eq3)
This review also includes intramolecular cycloadditions involving atoms other than carbon, such as oxygen in carbonyl groups (Eq 4). A number of intermolecular
E-', d
a-', d
(Eq 6 )
Several very limited reviews of the intramolecular photochemical cycloaddition of nonconjugated olefins
(Eq4)
373
WENDELL L. DILLING
374
have appeared recently in conjunction with their more common intermolecular counterparts (12, 52, 57, 62, 63,69,79,81,87,104). A discussion of the proposed mechanisms and excited states involved in these reactions is included where possible. The reactions are grouped according to structural types rather than mechanistic types since in a majority of the cases the mechanism is unknown or only partially known. The terms acyclic, monocyclic, bicyclic, etc., refer to that portion of the molecule containing the two reacting unsaturated linkages, not the molecule as a whole. The sources of radiation were generally medium- or high-pressure mercury arc lamps unless otherwise specified.
11. CYCLOADDITIONS INVOLVING CARBON 1,2 ADDITION A.
SCHEME I
+
Hg Hg*
+
hv
-
Hg*
---t
+
Hg
c*
I
c* c* DJ=* -+
decomposition
+
@ * + M + c + M I
ACYCLIC COMPOUNDS +
decomposition
1. Olejins
The mercury-sensitized isomerization of 1,5-hexadiene (I) in the vapor phase gave a mixture of the bicyclohexane (11), allylcyclopropane (111), and other free-radical-derived products (93) (Eq 7). The yield
+ I
I1
(Eq7) I11
of I1 increased gradually with increasing pressure in the range 0-200 mm, while the yield of I11 reached a niaximum a t a pressure of 28 nim of I and then decreased with increasing pressure. The yields of the other freeradical-derived products continuously decreased with increasing pressure. The quantum yield (a) for I11 at 28 mm was ca. 0.10, while that of 11was ca. 0.012 a t 200 mm. Two mechanisms were proposed for this reaction on the basis of the quantum yield dependence on the pressure (93). The first is shown in Scheme I. The second proposed mechanism involved three vibrational levels (or sets of levels), 2, 1, and 0, of the excited electronic state of I (see Scheme 11). Both mechanisms predict maximum quantum yields for I1 and I11 a t some pressure. That I1 did not show a maximum was attributed to the experimental difficulties of determining the maximum due to a combination of a relatively high (93). The procvalue of M,,, and a low value of krtx ess Hg* -+ Hg at 2637 A involves a triplet + singlet transition (e3P1-t SlS,), and therefore the excited state of I was probably a triplet. The formation of I1 rather than the isomeric bicyclo [2.2.0]hexane (IV), which
m IV
would have arisen from the alternate mode of addition
I11
6"Gp; -
decomposition
1
-+
free-radical products
P H O T O C H E M I C A L CYCLOADDITION
REACTIONS O F NOSCONJUG.4TED
SCHEME IV
of the double bonds of I, may reflect the instability of IV with respect to I1 (Scheme 111). The excited IV, if formed, may have decomposed to I or other products. SCHEME I11
Hg*T
+
375
OLEFINS
V
cs
r
l*
I
VI11
I1
0 a> A
‘..i,
inversion spin
~
VI
‘t IV
Mercury-sensitized photoisomerization of 1,6-heptadiene (V) at 3-65 mm pressure gave both of the possible intramolecular cycloaddition products, VI and VII, in addition to other products resulting from hydrogen migration, VI11 and IX, and free-radical processes (96) (Eq 8). The quantum yield for the disappearance
a & 6.a vapor ’iv,Hg phase
,
+
V
VI
VI1
SCHEME V
+*t
VI
VI1
+
Irradiation or myrcene (XI) a t X >220 mp gave a mixture of products (Eq 9), the major ones being 0pinene (XII) and the cyclobutene XI11 (32-35). A
VI I
Me
+
. 2 unidentified isomers
4-
L
VI11
IX
Hs -t free-radical products f liquid polymer (Eq 8) Y
J
54-68 %
95 70
of V was 0.6 a t 15 mm and 0.3 at 46 mm. The quantum yield for the formation of all isomers was 0.015 a t 2165 mm. At lower pressures the latter quantum yield decreased. The amount of VI increased with increasing pressure while that of VI1 appeared to reach a maximum a t ca. 15 mm. More of VI was formed than VII, the ratio being ca. 4: 1 a t the higher pressures. The mechanism for the formation of VI and VI1 is believed to be similar to that of I1 above, the relative amounts of VI or VI1 and VI11 depending on the pressure (96). A “hot” cyclobutane derivative (X) was proposed as the precursor to these isomers (96) (Scheme IV). In this case (Scheme V) the excited triplet of V closed in both directions.
+
other products (Eq9)
XIV
small amount of XIV has also been observed in this reaction (35, 37). I n contrast, irradiation of XI in the presence of a sensitizer (Eq 10) gave the isomeric bicyclo[2.l.l]hexane XIV as the only product (61). Ben-
hvsR2COsensitir hexane or benzene
XIV
XI
WENDELLL. DILLING
376
zophenone, &acetonaphthone, and fluorenone all act as sensitizers for this reaction; the quantum yield in the case of benzophenone was 0.05. The mechanism in the presence of a sensit,izer (Scheme VI) is believed to involve the transfer of triplet energy to the olefin X I and subsequent ring closure (61). It appears that the nonsensitized reaction (Scheme VII) (32-35) may proceed via the excited singlet ala*state of X I (61). SCHEME
RKQ
+
hv
XVII
CHO XVIII 9%
R*CO*T
+
+
&H
VI
XIX 18%
of the cis and trans forms of XVII is probably rapid compared with the cyclization (26). One mechanism proposed for the formation of XVIII involved the n -+ a* excited state of XVII mixed with some chargetransfer character (26). The high extinction coef325 mM) ficient (E 72) of the n + A* transition (A, of XVII was attributed to the mixing of the pure localized n + A* transition with a charge-transfer (CT) transition from the isolated double bond to the same antibonding ?r* orbital (Eq 13). This excited state
A XI
0 SCHEME VI1
XVII
XI -\+
OAc
I
(Eq13)
1
c.1.
XI1
Irradiation of neryl acetate (XV) in the presence of benzophenone gave the isomeric acetate XVI as 90% of the volatile product (16) (Eq 11).
etc.
may be either a singlet (26) or a triplet (11,26). Ring closure could then give XX or a zwitterionic form, followed by the second ring closure (15, 16, 26) (Eq 14).
- 0:
I
xv
\
XVI
a,p-Unsatusated Carbonyl Compounds A mixture of cis- and trans-citral (XVII) gave the cycloaddition product XVIII and a hydrogen migration product (XIX) on irradiation at the wavelength of the n + a* transition of XVII in either cyclohexane (25, 26) or ethanol (11) (Eq 12). The interconversion 2.
xx
XVIII
The alp-unsaturated ketone, carvone (XXI) , was converted to carvonecamphor XXII on irradiation through Pyrex (10, 14, 88, 89) (Eq 15). A more detailed study of this reaction using artificial light sources has shown that the product XXII was photolyzed t o the ester XXIII (or XXIV) during the reaction (66).
PHOTOCHEMICAL CYCLOADDITION
377
REACTIONS OF NONCONJUG-4TED OLEFINS SCHEMEVI11
25% EtOH, 6.5 months
XXI
XXII, 9.401,
Use of a black-light fluorescent lamp (maximum intensity at ca. 355 mp) gave larger amounts of the isomer XXII than did a high-pressure mercury arc lamp with a Pyrex filter. These results indicated that higher energy radiation was required for the destruction of the saturated ketone XXII than for its formation. Irradiation of 0.5% solutions of carvone (XXI) with the
XXII
XXVII and XXVIII, respectively, as the only distillable products (28) (Eq 16 and 17, respectively). IiY
black light
5 4 XXI
XXII
XXIII
XXIV
black-light source gave increasing amounts of the ester
XXIII after longer exposure times; larger amounts of the isomer XXII were formed with solvents of greater polarity (66) (see Table I). On a preparative scale, T.4BLE
Solvent
EtOH
%yoEtOH 51eOH HCONhlez MeCN CaHa EtzO Cyclohexane
1
Time, days
---% XXII
1 2
13
32
6
3 4 5 6
44 44 27 12
7 3 3 3 3 3 3 3
3 46 46 13 12 5 4 2
24 56 73 88 07 35
conversion---
XXIII
XXIV
..
+
The a,@-unsaturated ester XXIX was converted to the same mixture of isomeric cyclopentane derivatives XXX by either photosensitization with acetophenone or direct irradiation (8) (Eq 18). This reaction probably occurs via the triplet state of XXIX (8). I n a
e
E \
t OEt
hv, Pyrex cyclohexane, CGH,COMe or hv, silica
'
A 39
XX1.X
137; unknown
the ketone X X I I could be isolated in 35y0 yield by irradiation of XXI in 95% ethanol for 23 days (66). The formation of XXII presumably involves the n , r * state of the a,@-unsaturated carbonyl system (10) and may involve the triplet state based on other work with cyclic a,@-unsaturated ketones (41) (Scheme VIII). Two other alkenylcyclohexenones, XXV and XXVI (1% solutions), have been iiitramolecularly cyclized to
x x x , 55 %
similar manner the diacid XXXI and the diketoile XXXII were cyclized to the isomers XXXIII and XXXIV, respectively (8) (Eq 19 and 20, respectively). The dimethyl ester XXXV gave an all-trans dimer on irradiation with sunlight through glass (18, 29, 67, 97, 100). The dimer could have either structure XXXTI (57) or XXXVII. The half-closed djnier XXXVIII
WENDELLL. DILLING
378
.
o HO2CA
XXXI
I
C 0 XL
2
H
hv,sunlight
glass solidstate
XXXIII
O
XXXII
XXXIV
(57) was probably an intermediate in the formation of XXXVI or XXXVII (Eq 21). A MO2C
c
0
2
M
e
xxxv
hv, sunlight glass d i d state
XLI
XLII
LI was also isolated from XLIV and XLV (Scheme IX) . SCHEYE IX
0 hv
MeO$
H
Me
--80-20",Ar
0
The diethyl ester XXXIX was rather inert to irradiation compared with the dimethyl ester XXXV (100). An unidentified oil was formed on irradiation in the solid state and a ketone or aldehyde in solution (97).
XLIX hv XLV
XLVII
.1
+the Me
Me
"'aMe LI 0
0
/I
Et20
XLIV
XXXVIII
EtO&CH=CHCCH=CHCO:E
XLV
t
& hv
+
Me
XXXIX RCOR' or RCHO
The diacid X L was converted to the same carbon skeleton XLI or XLII as formed from XXV (67), presumably via XLIII (Eq 22). Duroquinone (XLIV) added twice to the olefins XLV and XLVI to give XLVII and XLVIII, probably via the open dimers XLIX and L (59). A 2 :1 adduct
Me
0 XLIV
Me
XLVI
L
XLVIII
PHOTOCHEMICAL CYCLOADDITION REACTIOXS OF NOKCONJUGATED OLEFIXS B.
SCHEME XI
MONOCYCLIC COMPOUNDS
I,
Olefins
The mercury-sensitized intraniolecular photocycloaddition of cis,cis-1,5-cyclooctadieIie (LII) in the gas phase gave a low yield of the tricyclooctane LIII in addition to other products (91, 94) (Eq 2 3 ) . The
-
LII
+
cj
LIII
LVII
I:?
A
03 LII
379
H
4- Mev-yMe 4
polymer (Eq 23)
+
1
-l
OEt H LIX
OEt OEt LIII, 28%
OEt
I
Mey-vMe
LVIII
LJV
u
H 1
2-301,
crossed olefin addition is again the preferredmode of ring closure; none of the isomeric tricyclooctane LV, which would result from c i s addition, was detccted (91).
1
MeC-CMe I
/
H
t
OH OEt LX
OEt
I
I
MeC-CMe I
OH
I
H
LX I
lytic action of cuprous chloride involved complexiiig with the excited state of LII (94) and is shown in Schenie XII.
L\
SCHEME XI1
Since the product LIII was derived froin the trans,tmns ihomer of LII, either both double bonds of LII had to twist in the excited state or the product LIII was derived from the small portion of the cis,trans isomer of LII present as an impurity in the LII (91). In either oaqe it appears that the tyans isomer of LV, LVI, could arise by formation of a 4 6 ring system, whereas formation of LIII requires a 5,s ring system. Again a triplet excited state was probably involved (Schemp X). It was pointed out that LV or LVI, ScaEJrE
H
s
4
LII
LIII
An alternate mechaiiisni has been proposed for the cuprous chloride promoted ring closure of LII to LIII (4). Under conditions similar to those noted above (using a high-pressure mercury arc source) the deuterated diene LXII gave the isomer LXIII containing less deuterium than the starting material (4) (see Eq 24). The undeuterated diene LII gave LIII in 39..575
LXII
on(’e formed in an excited state, niay not have been stable and may have decomposed before deactivation to their ground states could occur (91). In contrast to the gas phase photochemistry of LII, in solution in the presence of cuprous chloride reasonable yields of LIII were formed (92, 94) (Scheme XI). The quantum yield for the formation of LIII was ca. 0.1 based on the amount of LII. The quantum yield was too high (ca. 10) based on the amount of complex LVII which could be present. Therefore the diene LII must absorb most of the radiation. The solvent-derived products, LVIII-LXI, were alho formed in the absence of LII. A niechanism proposed for the rata-
2 . 3 7 , do 96.770 di 1.0% dz
LXIII, 24v0 14.470do 84.7% dl 0 . 9 % df
yield under these conditions. The calculated deuterium content of LXIII for the free-radical mechanism shown in Scheme XI11 was 13.4% do, 86.1% dl, and 0.5% dz (4). Irradiation of the tetraphenylcyclooctadiene (LXIV) was originally thought to give a compound analogous to LIII (106). However, a recent report (106a) showed that the products were diphenylacetylene and probably a 1 : 1 adduct of LXIV and isooctane (Eq 25). A dihydrophenanthrene of partial structure LXV uas also a possible product.
WENDELL L. DILLING
380 SCHEME XI11
L
LXII
gas phase (77) gave the same products plus another isomer, LXXIII (Eq 27). A reasonable primary process for this reaction is ?T + ir* excitation of the conjugated diene system since photons of this energy would be available under the reaction conditions. Scheme XIV was proposed to explain the products LXVII and
1
.1
SCHEMEXIV
FCHzMe
OD
+
LXVI
EtObHMe
r
+ LVIII-LXI
1
LXIII
Irradiation of 1,3,6-cyclooctatriene (LXVI) in pentane through quartz with a high-pressure mercury arc Ph
1
Ph
Q
hv. 2537 A c PhCECPh
isooctane
A
+
Ph Ph
LXIV
LXVIII LXIV-isooctane adduct
+
(Eq 25)
LXV
gave quantitatively a mixture of products, among which were the products of 1,2 addition, LXVII, and a combined 1,2-1,4 addition, LXVIII (116) (Eq 26).
LXVI
LXVII, 11%
LXX, 31%
LXXI, 20%
LXVIII, 8% LXIX, 16%
LXVIII (116). The question of whether a singlet or triplet excited state is involved is not yet settled (116). 2.
a#-Unsaturated Carbonyl Compounds
The dianhydride, byssochlamic acid (LXXIV), was converted to photobyssochlamic acid having either structure LXXV or LXXVI (3, 5) (Eq 28). Structure LXXV was favored based on the thermal stability exhibited by the photoproduct (3). 0 "rk0
I, v
LXXII, 13%
Another report indicated that similar irradiation of LXVI in either methanol solution (1%) (77, 78) or the
LXVI
LXVII, 18yo
LXVIII, 10%
LXXV
70%
C.
LXXVI
BICYCLIC COMPOUNDS
1. Bicyclo [2.2.1 Iheptadienes LXIX, 30%
LXX, 8 %
LXXI, 20 %
LXXII, 9 %
LXXIII, 5 %
Bicyclo [2.2.1Iheptadiene (LXXVII) was converted to quadricyclene (LXXVIII) by direct irradiation in ether solution (36) (Eq 29). The ultraviolet absorption spectrum of LXXVII in the vapor state showed a series of 17 bands a t 1985-2258 A with maximum intensity a t 2113 A (55, 107). A maximum was also ob-
381
PHOTOCHEMICAL CYCLOADDITION REACTIONS OF NONCONJUGATED OLEFINS SCHEME XV
LXXVII
&
LXXVIII, 67%
+
h"
- &; I .
LXXVII
served a t 188 mp (55). I n ethanol solution LXXVII had absorption maxima a t 205 mp ( E 2100), 214 (1480), and 220 (870) and a shoulder a t 230 (200) (107). Since nonconjugated olefins do not absorb radiation a t energies this low, there apparently is interaction between the two double bonds of LXXVII (36). Bicyclo[2.2.l]hept-2-ene (LXXIX) had an absorption maximum a t 195 mp in ethanol (107)) and significant maxima a t 1956 A ( E 5410), 2064 (3210), and 2077 (2940) in the vapor state (98). Molecular orbital calculations indicated a bond order of 0.12 or 0.50 for the C-2-C-6 bond depending on the model used and 1.50 for the C-2-C-3 bond in the lowest excited state of LXXVII, as shown in LXXX (107).
LXXIX
LXXX
Irradiation of LXXVII in the vapor phase at 4-31 mm and 25-50' with the 2537-A line of a low-pressure mercury arc failed to give any LXXVIII, the products being cyclopentadiene (LXXXI) (a = 0.50),acetylene (LXXXII) (a = 0.49), and toluene (LXXXIII) (a = 0.056) (75). 1,3,5-Cycloheptatriene was also formed at higher conversions (76). Addition of small amounts of oxygen (up to 36 mm with 10 mm of LXXVII) or hydrogen had no effect on the quantum yields. The lack of effect of oxygen indicated that only excited singlet states were involved (75). Addition of 650 mm of hydrogen or helium decreased the quantum yield slightly for LXXXI and LXXXII but had no effect on the quantum yield for LXXXIII. I n ether solution, in which the reaction is equivalent to that in the vapor phase at extremely high pressure, LXXVIII was formed along with LXXXI, LXXXII, and LXXXIII (75)(Eq 30). Based on these data the mechanism
&-A+ LXXVII
LXXVIII
Q LXXXI,
LXXXI
+
~i = 0.12
LXXXIII, a-0.042
shown in Scheme XV was proposed where H designates a higher vibrational level of an excited state, and L a lower vibrational level (75). The origin of LXXXIII
HCZCH LXXXII
h* p$ L
+
LXXVIII
may be a different vibrational level of the electronic state giving LXXXI, LXXXII, and LXXVIII, or a different electronic state entirely (75). The same transformation, LXXVII -P LXXVIII, has also been achieved in high yield by photosensitization using acetone, acetophenone, or benzophenone (53) (Eq 31). A more detailed examination of this reaction
LXXVII
LXXVIII,
N
95%
has shown that the process is reversible (54) (Eq 32). The relative amounts of LXXVII and LXXVIII dehv medium pressure H g arc
LXXVIII
LXXVll
pended on the triplet energies of the sensitizer. The sensitizers, R2C0, used are shown in Table I1 (54). TABLE I1 PHOTOSENSITIZED CYCLOADDITION OF BICYCLO [2.2.1]HEPTADIENE (LXXVII) % LXXVIII
% LXXVII LXXVIII after 6 hr
0 16 10 2
RaCO
kcal/mole
from LXXVII after 6 hr
CaH6COMe (CeHs)zCO 2-Naphthaldehyde (CeH&O)z
73.6 68.5 59.5 53.9
100 86 53 50
ET,
Me
LXXXII, @=0.14
h i 0+
from
With fluorenone (ET= 54 kcal/mole) as a sensitizer it appeared that the photostationary-state concentration of LXXVII would be >SO% in mixtures with LXXVIII (54). Side reactions were a serious complication in the study of these reactions. Side products acting as sensitizers apparently accounted for the large amount of LXXVIII at the end of the reactions. Two mechanisms
382
WENDELL L. DILLING
have been proposed for the LXXVII-LXXVIII interconversion (54). One involved excitation of either LXXVII or LXXVIII to triplets having approximately the same nuclear configuration in accordancewith the Franck-Condon principle (Scheme XVI). An alter-
LXXXV
LXXXVI
LXXVII gave the ring-closed isomers, XCI-XCIV, on irradiation in the presence of a sensitizer (not specified but probably acetone, acetophenone, or benzophenone) (99) (Eq 35).
SCHEME XVI
LXXVII
LXXVIII
nate mechanism involved the breaking of single bonds of LXXVIII to give a biradical (54) (Scheme XVII). SCHEME XVII
+
R~CO*T ---t
A 3 + t* *t
LXXVIII
RzCO
LXXXVII, x = c1 LXXXVIII, X = OH LXXXIX, X = OEt XC, X = p-BrCsH4SOa-
The bisbicycloheptadiene (XCV) gave the saturated isomer XCVI on direct irradiation (101) (Eq 36). The ultraviolet absorption spectrum of XCF’ was similar
LXXXIV
I
spin inversion
hv, quartz pentane, 12hr,
LXXVII
Presumably the reverse of this process could account for the LXXVII + LXXVIII conversion if LXXVII were converted to LXXXIV by energy transfer from R&O*T (Scheme XVIII). SCHEXE XVIII
f L /b +
RKO*T
- &I* f +
RzCO
t*
LXXVII
XCI, x = c1 XCII, x = OH XCIII, X = OEt XCIV, X = p-BrCt&SOa-
xcv
XCVI, 36%
to that of LXXVII (101). The diinethyldimethoxy derivative XCVII gave XCVIII in good conversion on irradiation in the presence of 2,4-dimethylbenzophenoneas EI sensitizer (46) (Eq 37). MeOyOMe
MeO,OMe
LXXXIV
I
spin
Me
inversion
hv, “BlacMite” 0,
Me XCVII
p-Me&HsCOCsHs pentane, -10’
XCVIII
The bicycloheptadienedicarboxylic acid (XCIX) has LXXVIII
The formation of LXXVIII could also be catalyzed or sensitized by cuprous chloride (95) (Eq 33). The exact function of the metal ion is not clear.
+ LXXVII
CUCl
-
XCIX
hv (2537 A)
0% 33)
Et20
LXXVIII, -70%
Several 7-substituted bicycloheptadienes have been converted to the corresponding 7-substituted quadricyclenes by irradiation. The 7-acetoxy derivative LXXXV gave LXXXVI on direct irradiation (74) (Eq 34). Several other derivatives, LXXXVII-XC, of
c,>47%
also been converted to the quadricyclene isomer C (30, 31) (Eq 38). The ultraviolet absorption spectrum of XCIX had a maximum at 243 mp ( E 5400) (31) [247 mp (e 5150) (%)I and end absorption, E 13,000 at200 mp in 95% ethanol (58). The low-energy absorption XCIX probably involves excitation to an electronic state in which the electrons of the nonconjugated double bond are involved. Once this excited state CI of XCIX
PHOTOCHEMICAL CYCLOADDITION REACTIONS OF NONCONJUGATED OLEFINS is reached it could easily pass to the closed form C (58) (Eq 39).
mMe hv, Pyrex
CVI
XCIX
383
Me
TziGz
+
I2 hr
j:&[
CI
-
Me
AM Me CXII, 58%
CXVII C02H C
2. Other Bicyclic Compounds The Dewar benzene CII was closed to the prismane CIII by irradiation a t 2537 A (108). The reverse reaction also occurred, being part of the over-all photochemical isomerization of the tri-t-butylbenzenes (108) (Eq 40). At the photostationary state CIII con-
hv, Pyrex
+
Me
96 hr
CVII
or f Me
e
d
p Me j
-I
L
CXVIII
CXIX
.1 CII
Ph
CIII
stituted 64.8% of the total reaction mixture and CII, 7.1%. A series of methylated naphthalenes CIV-CVIII in 3 4 molar excess were irradiated with diphenylacetylene (CIX) to give caged adducts CX-CXIV containing two cyclobutane rings (80) (Scheme XIX). These SCHEMEX I X
a
-
hv, Pyrex
-
Me +
hv, Pyrex
PhCsCPh CIX
CIV
CXIII, 44%
13 hr
CVIII
cvclohexane 120 hr
r
Ph
1
Ph
Me&:" Or
L
[Me
cxx
CXXI
4
Me
-
Ph \
hv, Pyrex
11 hr
cv
Me
Me r
Me
1
A; CXIV, 45%
-1
L
CXVI
CXI, 57%
products were probably formed via the cyclobutenes CXV-CXVII, CXVIII or CXIX, and CXX or CXXI, respectively (80). Attempted intramolecular cycloaddition of CXXII
WENDELLL. DILLING
304
and CXXIII by irradiation in the solid state or in ethyl acetate solution gave only tars or dimers (20).
0‘
cxxx cxxm
CXXII D.
TRICYCLIC COMPOUNDS
0
I. endo-Dicyclopentadienes
CXXXI, 89%
A number of endo-dicyclopentadiene derivatives have been converted photochemically to the closed cage The isomers, pentacyclo[5.3.0.0~~6.03~9.0S~*]decanes. parent hydrocarbon CXXIV was converted to CXXV by employing acetone as a sensitizer (84,85) (Eq 41). Pinacol (CXXVI), biacetonyl (CXXVII), a 1 :1 adduct CXXVIII of the diene CXXIV and acetone, and numerous other products were also formed (39). Under
4 CXXIV
hv’quartr
+
@+
OH O H
MezC-CMez I I
TABLEI11 ULTRAVIOLET ABSORPTION MAXIMAOF DICYCLOPENTADIENONE (CXXX) Solvent
* + **
Cyclohexane EtOH
220.3(7810) 227.0 (7730)
LSX, mfi (e)
CT 245 sh (900) 257 sh (720)
n
--c
**
342.8(37) 323.0 (3s)
absorption was attributed to charge transfer from the olefinic double bond to the a,p-unsaturated ketone chromophore (23) (Eq 43). However, since the reac-
Me,CO, 18hr
CXXV, 62 %
+ (MeCOCHz +r + CXXVII
CXXVI C13H180 (Eel 41) CXXVIII
the same conditions using cyclohexane as a solvent without acetone only a small amount of an unidentified brown oil formed (85). It has been postulated that this ring closure involves nonvertical excitation energy transfer to CXXIV (79). The same type of process has been proposed for the sensitized ring closure, LXXVII to LXXVIII (79). The ultraviolet absorption spectrum of CXXIV indicated that some type of interaction does occur between the double bonds in the excited state since CXXIV absorbed a t longer wavelengths than either bicycloheptane (LXXIX) or cyclopentene (CXXIX) , Dicyclopentadiene (CXXIV) absorbed weakly in the range 220-300 mp with increasing intensity toward shorter wavelengths; a26 =40 in hexane (72). Cyclopentene (CXXIX) absorption in the gas phase dropped to B =30 a t 211 mp and fell sharply in intensity above 210 mp (71). Bicycloheptene (LXXIX) absorption dropped sharply above 208 mp as noted above (98). I n the absence of interaction of the two double bonds, CXXIV should absorb a t wavelengths no longer than those observed for LXXIX or CXXIX. Dicyclopentadienone (CXXX) was smoothly converted to the cage isomer CXXXI on irradiation through Pyrex (17,23,27,51) (Eq42). Irradiation of CXXX in the crystalline state gave only a polymer (17). The ultraviolet absorption spectrum of CXXX is given in Table 111 (23). The charge-transfer (CT)
*
cxxx
(Es 43) -0’
tion proceeded readily through a Pyrex filter which effectively cuts off radiation at wavelengths less than 280 mp, this transition was probably not involved in the production of CXXXI. More likely the n + x* transition was the initiating process (Scheme XX). This SCHEME XX
cxxx
CXXXII
0 CXXXI
PHOTOCHEMICAL CYCLOADDITION REACTIONS OF NONCONJUGATED OLEFINS transition could account for the formation of CXXXI. Intermolecular attack of the excited state CXXXII would account for the formation of polymer. The above reaction, CXXX + CXXXI, also occurred on irradiation through quartz (38). The tetrachloro analog of CXXX, CXXXIII, gave an analogous product CXXXIV on irradiation (111, 112) (Eq 44).
CXLI
CXLIII
(Eq44)
CCla, -12 hr
0
CXLII
CXXXfV, 24 %
CXXXIII
reasoned that the failure to obtain the caged diketone CXLIV in the nonprotic solvents was due to the sp2
The dibromoketone ketal CXXXV gave CXXXVI in high yield (43) (Eq 45).
@
Br
Br
cxxxv
385
go
Br
0
CXLIV
character of the one-carbon bridge in CXLI while in CXLIII this bridge is sp3 (42). On the other hand, the perchloro diketone CXLV gave good yields of CXLVI on irradiation through quartz (40, 51) (Eq 49). The ult'rsviolet absorption
0
CXXXVI, 95y0
I n an analogous manner the octachloro ketone ketal CXXXVII gave the caged isomer CXXXVIII (45), while the ethylene ketal CXXXIX gave CXL (86) (Eq 46 and 47, respectively).
d'
0
CI
-
CXLV
CXXXVII
CXLVI, 73434%
spectrum of CXLV showed an n P* transition above 300 mp (Table IV) (39, 64, 105). Irradiation of CXLV through Pyrex (A >-280 mp) in methylene chloride solution gave CXLVI efficiently, -90% (39). Therefore the n , r * state (singlet or triplet) of CXLV is probably the reactive state.
CXXXVIII
TABLEI V ABSORPTION ?VIAXIMA OF OCTACHLORODICYCLOPENTADIENEDIONE (CXLV) (39) Loax, mr (4 ULTRAVIOLET
hv, Pyrex ___t
CI
-CI CI
CI
(Eq 47)
Solvent
Hexane
r+
7T*
254.5 (8800)
0 CXXXIX
CXL, 100%
Attempted ring closure of the dibromodione CXLI in benzene or methylene chloride gave polymer (42). I n methanol containing hydrogen chloride the caged isomer CXLII was readily formed (42) (Eq 48). It was
MeCN
258
(9400)
n
-.+
316 329 949.6 359.5 378
r*
(83) (101) (104) (77)
(32)
-310 sh (110) 323 (107) 936 (108) 351 (82) 367 sh (36)
WENDELLL. DILLISG
386
A compound closely related to dicyclopentadiene (CXXIV), dicyclohexadiene (CXLVII), was not closed to the caged isomer CXLVIII on irradiation in acetone or benzene solution with sensitizers such as benzophenone, 0-acetonaphthone, benz [ulanthracene, eosin, or 9,lO-dibromoanthracene (102) (Eq 50).
Several closely related compounds CLI-CLV were also converted to the cage isomers CLVI-CLX on irradiation, probably via similar mechanisms (20, 21, 23) (Scheme XXII). SCHEME XXII
EtOAc, hv,Pyrex Nz, 12 hr
, $f0
0 CXLVII
CXLVIII
0
CLI
CLVI, 40%
The ezo isomer of CXLVII was also unaffected under the same conditions (102). 2. Quinone Adducts A number of Diels-Alder adducts of cyclic dienes and quinones which have the endo configuration have been converted to the caged isomers by irradiation. The adduct CXLIX of cyclopentadiene and quinone gave CL on irradiation either in solution or the solid state (20, 21, 23) (Eq 51). The ultraviolet spectrum of
9 0
solid state, 20 hr, 60% yield
CLII
CLVII
petr ether, hv,Pyrex Nz, 3 hr
, $o
0 6
1
80 hr, 80% yield
CXLIX
, &o
(Eq ~ 51)
~
~
0
~
~
CLIII
0 CL
CLVIII, 4 w 0
CXLIX is given in Table V (23). Irradiation of CXLIX TABLE V ULTR.4VIOLET ABSORPTION MAXIMA O F endo-TxIcYcLo [6.2. 1.02~7]LTNDECA-4,9-DIENE-3,6-DIONE (CXLIX) 7
Solvent
Cyclohexane
226 238 226 239
EtOH a
** (12,900) sh (6,800) (13,000) sh (8400)
r+
(4 CT
-.
CLIX, 54%
CLIV
Amax, m p
n
+ u*
278 (248) 385 fa (58) 284 (254)
hu, Pyrex
374 f (63)
EtOAc, Cob85 hr
Fine structure.
through a filter which cut out all wavelengths shorter than415 mp gave a small amount of CL (20). Therefore the initial excitation of CXLIX in the photochemical reaction was probably n + T* (20). The mechanism may be pictured as shown in Scheme XXI. SCHEME XXI
L
CXLIX
.I
-
-
h0-c$30- Q0j 0
CL
0
L
0
- - I
CLV
CLX, 36%
Several quinone adducts of formal cyclohexadienes CLXI-CLXIII have been cyclized photochemically (20, 21, 102) (Scheme XXIII). The quantum yield for the formation of CLXIV was reported to be high (102). Under conditions similar to those shown above, compounds CLXVII-CLXIX failed to give intramolecularly cyclized products, giving instead dimers or tar (23).
dpoL2p-p 0 CLXVII
0 CLXVIII
+& 0 CLXIX
The caged adducts of duroquinone (XLIV), CLXX and CLXXI, were formed on irradiation of XLIV in
CYCLOADDITION REACTIONS O F NOSCONJUGATED OLEFINS
PHOTOCHEMICAL
387
SCHEME XXIII hv, Pyrex
+l
f 40
Et10 or solid state, 1-2 hr, 100% yield(l02) hv, Pyrex EtOAc, 10 hr, 80% yield (20)
hv, Pyrex solid state, 90hr, 90% yield (20)
0
0
CLXIV
CLXI hv, Pyrex EtOAc, 2.5 hr, 90% yield (20) hv, Pyrex
+
*:
solid state, 60 hr, 16%yield (20) L
CLXII CI
c+' 7J
0
CLXXIII
CLXV
CLXX, 15% (59)
CI
c'2$o
hv, Pyrex EtOAc, Cog, 9 hr
Me Me
CLXXV, 33%
0
CLtXIII
CLXVI, 35% (20)
the presence of the olefins LXXXI and CLXXII (59, 82). The half-closed adducts CLXXIII and CLXXIV would be likely intermediates (Scheme XXIV). From LXXXI the 1,4-addition product CLXXV was also isolated.
+
XLIV
CLXXII CLXXIV
me e&:
3. Quinone Dimers
I n this series of reactions some of the half-closed dimers have been isolated while others have not. The cis half-closed dimers gave the caged compounds readily on further irradiation. A caged dimer CLXXVI of p benzoquinone (CLXXVII) was obtained in low yield in a number of attempts to add CLXXVII to other unsaturated centers (9). The half-closed dimer CLXXVI11 of CLXXVII may be an intermediate in the formation of CLXXVI (Eq 52). Another report indi7
0 CLXXVII
L
---Me
U
CLXXI, 37%
2,3-dimethyl-p-benzoquinone have been reported, each of which gave a caged isomer, CLXXXI and CLXXXII, on further irradiation (6, 18, 44) (Eq 53 and 54, respectively). The question whether CLXXIX and C L S X X
-
(g) O H
H o
-1
CLXXVIII
0
IP,
Me
0
0
cated that CLXXVII gave only polymer and hydroquinone on irradiation (18). Two different open dimers, CLXXIX and CLXXX, of
H H
Me
sunlight
Me#
Me
orHgarc,solidstste
Me
0
0
0
CLXXIX
mMe 0
0
O M ~ H O
Me hv, sunlight
I
0 CLXXVI, 0.2%
0
0
Me H
0
CLXXX
Me-
0
(Eq54)
0
CLXXXII (44)
(and CLXXXI and CLXXXII) are really different has not been resolved (18). 2,BDimethylquinone CLXXXIII also gave a cage dimer of uncertain structure, CLXXXIV or CLXXXV,
WENDELL L. DILLING
388
in addition to the oxetane CLXXXVI (6, 18, 22, 24) (Eq 55). The half-cage dimers CLXXXVII and CL0 hv, sunlight or Hgarc,solidstate,4-8hr
Me 0 CLXXXIII
0
0
CXCIV
cxcv
The products and their stereochemistry of irradiations carried out in the solid st'ate appear to be highly dependent on the crystal structure in the solid.
4.
Pyrone Dimers Dimethylpyrone (CXCVI) gave a caged dimer CXCVII on irradiation in solution or the solid state (47, 70, 114, 115) (Eq 57). The half-closed dimer CXCV-
Me-
2%
CLXXXIV
CLXXXV
hw(sun1amp) r solid state, 3 weeks, 30% yield (115)
Me
J(l
hv(sunlamp), quartz
Me
CXCVI
t
H20,HOAC, EtOH, Or CeHs. 3 weeks,
